CASES OF LISTERIOSIS ARE BECOMING MORE COMMON - ARE READY-TO-EAT FOODS TO BLAME? - BFR
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
www.bfr.bund.de DOI 10.17590/20210311-115317 Cases of listeriosis are becoming more common - are ready-to-eat foods to blame? BfR Opinion No 004/2021 issued 2 February 2021 Over the last few years, cases of listeriosis in humans have been rising in Germany and across the EU. Food contaminated with Listeria monocytogenes has been the primary cause of these illnesses: although rare in the general population, infection with this pathogen can lead to fatal blood poisoning, encephalitis and meningitis. Pregnant women, babies, the el- derly and those who have an immune system weakened by pre-existing medical conditions or medication are all at greater risk of contracting a case of listeriosis. In pregnant women, who often have only flu-like symptoms or none at all, the infection can cause a premature birth or miscarriage. Listeriosis is not usually a health hazard for healthy adults: in most cases, the infection proceeds without any symptoms and therefore goes undetected. Over the last few years, ready-to-eat food products were the foods most often contaminated with Listeria monocytogenes. Listeria are both widespread and persistent in the environment. As a result, it is able to contaminate foodstuffs while these are being produced and pro- cessed. Although temperatures of 30 to 37 °C present ideal growth conditions for the bacte- rium, it can also reproduce slowly even in a refrigerator. If storage times are sufficiently long, colonies may grow to a size that represents a human health hazard. In light of the above, the German Federal Institute for Risk Assessment (BfR) has assessed 40 ready-to-eat foods capable of harbouring listeria to determine whether these have been consumed more frequently in recent years and whether such products were contaminated with Listeria monocytogenes. To do so, the BfR analysed data from market research com- pany GfK SE and results from food monitoring in Germany. A trend towards greater consumption was found for 14 foods from various categories, includ- ing pre-cut fruit salads, various ready meals that are consumed without first being (re- )heated, heat-treated fish and poultry products, deli salads, and fresh sprouted seeds and vegetables. During the last 10 years, ready-to-eat fish products such as cold- or hot-smoked fish and marinated (gravlax) fish were more frequently found to be contaminated with Listeria mono- cytogenes, as were meat products - especially raw (pork and beef steak) minced meat and spreadable raw sausage products such as pâtés. While Listeria monocytogenes was less commonly found in heat-treated meat products, concentrations of the pathogen in these foods were often too high. Occasional contamination with Listeria monocytogenes was also found in deli salads, cheese made from cow’s milk, speciality pastries and ready meals. However, the available data cannot be used to establish the increased consumption of such foods contaminated with Listeria monocytogenes as a contributory factor to the rise in cases of listeriosis in Germany with any certainty. This would require targeted surveys of consumption patterns in sensitive population subgroups and supplementary data for ready- to-eat foods not investigated sufficiently to date. Other reasons for the rise in cases of listeri- osis in Germany and the EU could include an increase in the number of susceptible individu- als and improvements in the systems used to report these kinds of illnesses. Seite 1 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de The BfR has also assessed a number of innovative technical methods for the production, preservation and distribution of ready-to-eat foods designed with the aim of reducing the inci- dence of listeria. The BfR has also published a consumer guide to avoiding food-borne listeriosis infections (see https://www.bfr.bund.de/cm/350/verbrauchertipps-schutz-vor-lebensmittelinfektionen- mit-listerien.pdf). For people at greater risk of contracting listeriosis, special rules apply for kitchen hygiene and for the consumption of certain kinds of foods. Above all, foods that are not (re-)heated before consumption should be prepared at home from fresh ingredients and consumed as soon as possible. Foods stored temporarily in the refrigerator should be con- sumed within two to three days. Subject of the assessment In recent years, the incidence of listeriosis in humans has been rising both in Germany and across the EU. Transmission to humans occurs primarily as a result of the consumption of foods contaminated with Listeria monocytogenes. As a result of recent listeriosis outbreaks in Germany caused by ready-to-eat foods, concerns have been expressed as to whether the hygiene standards for ready-to-eat foods that are intended to minimise the risk of infection are sufficient in the specific case of Listeria monocytogenes. In light of the above, the German Federal Institute for Risk Assessment (BfR) has addressed the following questions: 1. Over the last 10 years, is there evidence pointing to an increase or decrease in the con- sumption of ready-to-eat foods that promote the growth of Listeria monocytogenes or which could be contaminated with Listeria monocytogenes? 2. Which types of ready-to-eat food (e.g. of animal or plant origin) were most frequently found to be contaminated with Listeria monocytogenes during the last 10 years? 3. Is there evidence to support the hypothesis that the increased production of ready-to-eat foods and general growth in the market for ready-to-eat foods as part of the food retail sector has led to an increase in cases of listeriosis in humans? 4. Which innovative technological methods for ready-to-eat foods are suitable for prevent- ing the occurrence and growth of Listeria monocytogenes during production, handling and distribution? 5. Can additional risk communication measures be taken (and if so, which ones) to influ- ence the frequency of occurrence of infections involving Listeria monocytogenes? In the following sections, the BfR proceeds to assess these questions in the same order as introduced above. The answers given to questions 1 and 4 have been approved jointly by the Max Rubner Institute (MRI). Seite 2 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de 1. Response to the following question: Over the last 10 years, is there evidence point- ing to an increase or decrease in the consumption of ready-to-eat foods that pro- mote the growth of Listeria monocytogenes or which could be contaminated with Listeria monocytogenes? 1.1 Underlying data To answer the question as outlined above, the underlying data need to meet the following cri- teria: (1) Data are available at a level of aggregation that enables a distinction to be made on the basis of methods of production or processing that are relevant for contamination with Lis- teria monocytogenes (e.g. whether the food in question is a freshly-prepared or ready-to- eat, pre-packaged salad). (2) Data are available from at least two survey periods that ideally represent a trend line ex- tending over 10 years. (3) Data from both survey periods are comparable in terms of methodology and level of detail. Consumption data To answer research questions with relevance for consumption, the BfR can draw on repre- sentative consumption surveys conducted in children and adults from Germany (VELS [1], ESKIMO [2] and NVSII [3, 4]). For the following reasons, however, these data are not suita- ble for answering the present question in its full complexity: - The level of detail of the survey questions in the consumption studies does not provide adequate information about food preparation (e.g. distinction between fresh and ready-to- eat). - Data are also only available for a single point in time, which prevents the construction of a time series. Following a consultation with the MRI concerning the option of using National Nutrition Moni- toring (NEMONIT) data, it transpired that, while the follow-up study for NVA II for adults aged between 14 and 80 does permit the consideration of an additional period between 2008 and 2015, both the level of detail and sample size in this study would preclude their use in provid- ing a valid answer to the above question. In a NVSII subsample, approximately 2,000 individ- uals were also surveyed twice yearly about their food consumption using 24-hour recall. As data aggregation was performed for the evaluation at the food group level, the level of granu- larity no longer permits the consideration of individual foods. While a statement about cheese can be made, this is not possible for types of cheese such as soft cheese made from cow’s milk. Nor was the necessary level of detail provided about processing or the type of prepara- tion involved, such as whether fruit was pre-cut or salad was pre-packaged. The number of days surveyed is also too low to cover all of the foodstuffs eaten less regularly by any one individual. Such foodstuffs include fish, sprouted vegetables and raw milk cheese. In light of these problems, market data was consulted to obtain a valid answer to the ques- tion. While these data cannot substantiate statements about individual consumption quanti- ties/frequencies, they are nonetheless considered to be a current and adequately detailed dataset in this context. ► Market data In the preparation of answers to this question, survey data was used from the household and fresh food panel maintained by GfK SE. Seite 3 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de This household panel is a representative sample of households throughout Germany, whose purchasing patterns are recorded on a regular basis. The panel consists of a gross sample of 30,000 households, which record their purchases using a barcode scanner for transfer to GfK SE (to create the ‘household panel’ for foods such as cheese and dairy products, deli salads, ready meals, baked goods, etc.). Foods without an EAN barcode are either scanned using a code book or entered manually via an input screen. A subsample of 13,000 house- holds is also taken from the household panel, to form the ‘Fresh Food Panel’ (e.g. for fresh meat and fresh fish as well as fresh vegetables and fruit). The representativeness of the household panel and fresh food panel is ensured by making comparisons with the micro-cen- sus conducted by the German Federal Statistical Office. The use of market data offers a key advantage: the large sample size means that valid data can be collected about the increased or decreased consumption of products throughout Ger- many. Since the methodology remains the same, an exact comparison can also be made be- tween two different timespans. Barcode scanning also ensures a high level of accuracy since respondents are not asked to tax their memory about foods that they have consumed. In ad- dition, the barcodes can also be used to distinguish between different kinds of product pro- cessing and packaging, which allows information to be obtained that is of relevance for an- swering the above question (such as whether the fruit is pre-cut, for example). Food selection The term ‘ready-to-eat food’ is defined by EU Regulation (EC) No 2073/2005 of the Commis- sion of 15 November 2005 on microbiological criteria for foodstuffs as follows: “…food in- tended by the producer or the manufacturer for direct human consumption without the need for cooking or other processing effective to eliminate or reduce to an acceptable level micro- organisms of concern”. This definition includes both conventional ready meals as well as other kinds of products that can be consumed without further processing (e.g. fresh fruit, fresh vegetables). Since this definition encompasses a broad range of products, the selection of foods to be considered was restricted by applying the following two criteria: 1) Presence of relevant prevalence figures for Listeria monocytogenes from 2008 to 2017 2) Matrix or type of processing that promotes the growth of Listeria monocytogenes. Accordingly, 40 relevant foods were identified from the following groups: meat products, fish products, dairy products, baked goods, deli salads, ready meals, dried spices, salads and vegetables, and fruit (table 1). To enable an easier interpretation of the data, these foods are introduced briefly below. The selected foods include all types of ready-to-eat meat products, including heat-treated, dried and raw foods. Products are differentiated by animal species in each case. The heat- treated products include roasted or fried, stewed and cooked meat products, such as pork sausage, luncheon meat and cooked ham as well as pre-fried rissoles. The category ‘Other preserved meat products’ includes products such as raw sausage, raw ham or dried meat (jerky). Raw meat is minced (ground) into a ready-to-eat form and sometimes seasoned (steak tartare or pork mince with salt, pepper, onions). Minced meat is also included in this group, since it can be assumed that consumers may eat the product raw, may also taste the meat during preparation or may eat it without it being fully cooked through. Ready-to-eat fish products includes heat-treated products such as pre-cooked prawns, shrimp and crayfish, marinated products such as rollmops and fried herring or herring snacks Seite 4 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de (some of which reuse the marinade for cooking), as well as hot- or cold-smoked and gravlax products such as smoked trout or mackerel, and hot-/cold-smoked salmon and gravlax salmon. The dairy product group includes cheese made from the milk of various animal species as well as other dairy products. Cheese is differentiated based on soft or semi-hard types (such as brie, camembert or ‘Butterkäse’) and all other types of cheese. Dairy products excluding cheese include all other types of products such as yoghurt, curd cheese and buttermilk. The category of ‘Pastries’ includes not only the various Danish pastries eaten in Germany, featuring blancmange or cherry fillings, etc., but all other baked goods such as biscuit dough/yeast dough pastries and fruit cakes. With deli salads, specific recipe categories are used to distinguish between products - i.e. whether they contain fish, poultry, vegetables, egg or meat. The ‘Other’ section includes products using recipes for cheese or pasta salads, soy/tofu salads, festive salads and speci- ality salads from different cultures. This last category includes couscous, bulgur wheat, tab- bouleh and quinoa salads. The ‘Ready meal’ product group is composed of products that can be consumed without needing to be (re-)heated as well as products that need to be (re-)heated before consump- tion. Ready meals intended for consumption without further heating include complete meals, such as sushi, as well as sandwiches, filled rolls, bagels and baguettes, pasties, and ready- to-eat salads/bowls. Ready meals that need to be re-heated before consumption are further subdivided into non-refrigerated, chilled and frozen ready meals. The category of non-refrig- erated ready meals includes traditional dried products such as dried noodle or soup dishes. Chilled meals include refrigerated pizzas, soups or quiches. The last category of frozen ready meals includes frozen pizzas, other frozen meals (e.g. lasagne bolognese, nasi goreng, etc.) and frozen snacks (e.g. pizza baguettes, spring rolls, mozzarella sticks, pizza burgers, etc.). Although dried products and frozen products are viewed as less critical for contamination with Listeria monocytogenes, these products are listed for the sake of com- pleteness. It is also possible for a consumer to fail to heat through a technically ‘ready-to-eat’ product, only warming it up to a temperature that is merely comfortable for consumption: this particular case would constitute a microbiological hygiene risk. It should also be emphasised that the out-of-home consumption of ready meals could not be accounted for: accordingly, the evaluation does not include any ‘to go’ foods such as a sandwich bought at a rail station bake shop or other meals eaten at a snack bar, canteen or restaurant. Dried spices includes both individual herbs and spices as well as mixed spices and herb-fla- voured salts as a single item in the table. Salad and other vegetables are grouped into lettuce, leafy vegetables, vegetables that can also be consumed raw (e.g. tomatoes, cucumbers, peppers, etc.), and fresh sprouted vege- tables or seeds. A special category in the table is reserved for pre-cut vegetables and let- tuce. This category covers vegetables and lettuce that are already pre-cut into bite-sized por- tions and then packaged. This category constitutes a particular risk for contamination, since chopping the vegetables increases their surface area, thereby promoting proliferation and contamination. The ‘Fruit’ category covers the general fruit market. As with the previous section, this section also distinguishes between products that have been chopped into bite-sized pieces and Seite 5 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de packaged accordingly; these are considered separately under the category ‘Mixed fruit salad/(pre-cut) fruit’. Preserves were not included in the survey, since contamination of the product after heating in the can or jar is not to be expected. No data were available for certain items in the table. This affects the differentiation of dairy products (pasteurised/not pasteurised), the differentia- tion of smoked fish (hot-/cold-smoked), and the differentiation of pastries (with/without a cream filling). Time frame A survey extending over 10 years is not possible since the maximum retention period for market data at GfK SE is only five years. As a result, the evaluation extended only over the periods of September 2012 to August 2013 (survey period 1) and September 2017 to August 2018 (survey period 2). In each case, the data were analysed for a period of 12 months, so as to also account for seasonal variation during the year. Purchase volumes and buyer households The data as presented include sales volumes quoted in tonnes (t) and the percentage differ- ence in sales volumes between the two survey periods. In addition, the percentage differ- ence between buyer households between the two survey periods is also specified. As one example, the figure for households buying heat-treated meat products made from beef was 10% higher in 2017/18 when compared with 2012/13. It should also be noted that the num- ber of households in Germany also increased by 3.7% in the time between the two survey periods. In recent years, the structure of households has also undergone changes: as a re- sult, around 42% of households in Germany were only one-person households in 2017. A comparison of sales volumes across both survey periods provides evidence of a product being purchased in greater or lesser quantities on average by German households. From this, an indirect conclusion can be drawn as to whether the product was therefore being con- sumed more or less frequently. Accounting for the number of buyer households provides indirect information about whether the respective product is being consumed by more or fewer individuals, as well as the overall increase/reduction in consumption. This further substantiates the reliability of data on sales volumes, since portion and product sizes have been subject to systematic changes over the periods considered, and this could therefore influence the figures for sales volumes. If a product is being bought more often while simultaneously being consumed by more house- holds, then more frequent consumption overall is very likely. 1.2 Results Table 1 lists all of the foods selected, including sales volume in tonnes (t), as well as the dif- ference in this volume between survey periods (%) and the difference between buyer house- holds between survey periods (%). The items shown with a grey background can be under- stood as the ‘parent group’ for the categories shown below them - their sales volume figures represent the total for these foods. If only these parent groups are considered, then changes within the five-year period are small, with a difference in sales volume between −6% and +6%, and −2.7% and +0.3% in the number of buyer households. Consideration of the individual subcategories reveals a hetero- geneous picture, however. Across all of the foods considered, ‘Heat-treated meat products from other meat, excluding poultry’ (e.g. game) shows the strongest decline in sales volume Seite 6 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de at −43% (and −30% buyer households). ‘Mixed fruit salad/pre-cut fruit’ shows the strongest growth at +242% (+53% buyer households). When interpreting these data, however, one should remember that these products each make up a small proportion of the overall sales market. The sales volume for ‘Fruit salad/pre-cut fruit’ accounts for only 0.4% of the total sales volume of fruit, for example. Figures 1 and 2 provide an overview of the percentage changes in the two parameters con- sidered, namely sales volume and buyer households, over the two survey periods. These changes reflect the heterogeneity in the trends for the types of food considered. If a parallel rise in sales volume and buyer households is taken as an underlying criterion, then this indi- cates more frequent consumption for 14 of the 40 foods investigated. In the same way, 5 of 40 foods display a decline in frequency. These are listed separately in Table 2. All foods are included that exhibit a difference >1% in either of the two parameters. All other foods either show changes of ≤1% or exhibit opposing trend trajectories for sales volume and buyer households. A good example of the latter case is cheese made from buffalo milk. While sales volume rose by 39%, the proportion of buyer households fell by 9%: accordingly, larger quantities are probably being consumed by fewer households. Table 1: Selected ready-to-eat foods that have the potential to be contaminated with Listeria monocytogenes. Figures for sales volume in tonnes (t), percentage change in sales volume (%), and percentage change in buyer households (%) between the two survey periods of 2012/13 and 2017/18 (GfK SE, Consumer Panels & Services, 2018). Sales (t) % diffe- Survey pe- Survey pe- % diffe- rence riod 1 riod 2 rence Food buyer (9/2012 - (9/2017 - sales house- 8/2013) 8/2018) volume holds Heat-treated meat products 1,210,616 1,172,124 -3 -1 Poultry (chicken, turkey) 53,667 68,226 27 3 Poultry (other domestic poultry) 79,594 85,662 8 -5 Beef 19,285 21,919 14 10 Pork 1,048,946 991,149 -6 -1 Other meat (excluding poultry) 9,124 5,168 -43 -30 Other preserved meat products 320,718 317,816 -1 -1 Beef 1,700 2,990 76 37 Pork 318,533 314,376 -1 -1 Other meat (excluding poultry) 485 450 -7 -9 Raw meat 294,978 302,118 2 -2 Steak tartare, pork mince with seasoning 63,566 60,154 -5 2 Minced meat (overall) 231,412 241,963 5 -2 Fish produccts 105,667 99,408 -6 -3 Fish, heat-treated 4,637 7,006 51 23 Fish, marinated 51,812 52,412 1 -6 Fish, hot-/cold-smoked or gravlax 49,219 39,990 -19 -7 Cheese and dairy products 4,840,594 4,851,023 0.2 −1 Soft cheese/semi-hard cheese - cow’s milk 168,772 177,932 5 −0.3 Soft cheese/semi-hard cheese - goat’s milk 3,632 3,402 -6 5 Soft cheese/semi-hard cheese - sheep’s milk 251 307 22 13 Other cheese - cow’s milk 582,187 627,456 8 −0.2 Other cheese - buffalo milk 842 1,169 39 -9 Seite 7 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de Sales (t) % diffe- Survey pe- Survey pe- % diffe- rence riod 1 riod 2 rence Food buyer (9/2012 - (9/2017 - sales house- 8/2013) 8/2018) volume holds Other cheese - goat’s milk 8,734 8,545 -2 6 Other cheese - sheep’s milk 11,582 15,438 33 13 Dairy products (excl. cheese) 4,064,594 4,016,773 -1 -1 Pastries 149,566 158,740 6 -3 Deli salads 179,423 186,075 4 -1 Deli salads - with meat 52,485 51,035 -3 -5 Deli salads - with fish 19,076 19,295 1 -7 Deli salads - with poultry 7,103 7,259 2 -8 Deli salads - with vegetables 81,227 82,969 2 1 Deli salads - with egg 4,639 5,638 22 7 Deli salads - other 14,893 19,878 33 14 Ready meals 668,562 708,770 6 −0.4 Ready meals consumed without further heating 25,257 45,965 82 23 Ready meals consumed with further heating, 251,193 234,288 -7 -3 non-refrigerated Ready meals consumed with further heating, 75,919 96,123 27 3 chilled Ready meals consumed with further heating, 316,193 332,394 5 -1 frozen Spices, dried 16,468 16,225 -1 -2 Lettuce and vegetables 2,271,037 2,320,519 2 0.2 Pre-cut vegetables and lettuce 26,035 30,075 16 8 Lettuce 236,445 225,376 -5 -1 Leafy vegetables 27,562 31,771 15 16 Other fresh vegetables, excluding rhubarb 1,980,324 2,032,327 3 0.3 Sprouted vegetables and sprouted seeds 672 970 44 26 (fresh) Fruit 3,504,792 3,616,834 3 0.3 Fresh fruit, including rhubarb 3,500,961 3,603,722 3 0.3 Mixed fruit salad/pre-cut fruit 3,832 13,112 242 53 Seite 8 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de Figure 1: Percentage difference (%) in sales volume and buyer households as a comparison between the two survey periods of 2012/13 and 2017/18 (GfK SE, Consumer Panels & Services, 2018). Seite 9 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de 24 Figure 2: Percentage difference (%) in sales volume and buyer households as a comparison between the two survey periods of 2012/13 and 2017/18 (GfK SE, Consumer Panels & Services, 2018). [continued] Seite 10 von 55
www.bfr.bund.de Table 2 lists all foods that fulfil the criterion of a parallel rise/fall in sales volume and buyer households, sorted in descending order by percentage change in sales volume (%). The largest percentage changes, with increases in sales volume of over 50%, can be seen in ‘Mixed fruit salad/pre-cut fruit’ (242%), ‘Ready meals consumed without further heating’ (82%), ‘Other preserved meat products - beef’ (76%) and ‘Fish products, heat-treated’ (51%). Noteworthy is the fact that the top two foods originate in the conventional to-go segment, i.e. these foods can be consumed immediately after purchase (e.g. pre-cut fruit salad or a sand- wich purchased from the supermarket). It should also be mentioned that these figures are probably underestimates. Firstly, these products are often forgotten as a result of the imme- diate consumption (packaging is thrown away and not then scanned in at home). Secondly, identical products that are offered for immediate out-of-home consumption are not accounted for in the present figures (i.e. a filled roll bought from the baker at the station or a sushi-to-go box from a snack bar). Amounting to around 46,000 tonnes, ready meals sold for immediate consumption represent a large and highly diverse market. As a result of data aggregation, however, no further differentiation is possible into individual products. All of the other foods listed in Table 2 display sales growth figures ranging from 44% (sprouted vegetables) to 14% (heat-treated meat products made from beef). This also in- cludes a variety of deli salads, sheep’s cheese, heat-treated meat products made from poul- try and chilled ready meals, as well as pre-cut vegetables and lettuce, and leafy vegetables. A look at foods exhibiting declining trends in sales figures and buyer households reveals that products made from ‘other meat excluding poultry’ (e.g. game) are the primary types for which demand has fallen (−43% and −7%, respectively). Sales volumes for hot- and cold- smoked or gravlax fish products have also fallen significantly (by 19%). In this case, how- ever, no differentiation is possible between smoked fish types: this reduction may have af- fected both or only one type of smoked products. Lastly, conventional dried ready meals such as dried noodles and soups (‘Ready meals consumed with further heating, non-refriger- ated’) and deli salads containing meat were also purchased in smaller volumes (−7% and −3%, respectively). Seite 11 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de Table 2: Ready-to-eat foods with the potential for contamination with Listeria monocytogenes, for which a greater [↑] or lesser [↓] frequency of consumption is likely in the period 2017/18 compared with the period 2012/13. Sorted in descending order by difference in sales volume (%). Sales (t) % diffe- Survey pe- Survey pe- % diffe- rence riod 1 riod 2 rence buyer Trend Food sales vo- house- (9/2012 - (9/2017 - 8/2013) 8/2018) lume holds Mixed fruit salad/pre-cut fruit 3,832 13,112 242 53 ↑ Ready meals consumed without further 25,257 45,965 82 23 ↑ heating Other preserved meat products 1,700 2,990 76 37 ↑ - beef Fish products, heat-treated 4,637 7,006 51 23 ↑ Sprouted vegetables and sprouted 672 970 44 26 ↑ seeds (fresh) Deli salads - other 14,893 19,878 33 14 ↑ Other cheese - sheep’s milk 11,582 15,438 33 13 ↑ Heat-treated meat products 53,667 68,226 27 3 ↑ - poultry (chicken, turkey) Ready meals consumed with further 75,919 96,123 27 3 ↑ heating, chilled Soft cheese/semi-hard cheese - 251 307 22 13 ↑ sheep’s milk Deli salads - with egg 4,639 5,638 22 7 ↑ Pre-cut vegetables and lettuce 26,035 30,075 16 8 ↑ Leafy vegetables 27,562 31,771 15 16 ↑ Heat-treated meat products 19,285 21,919 14 10 ↑ - beef Heat-treated meat products 9,124 5,168 -43 -30 ↓ - other meat (excluding poultry) Fish products 49,219 39,990 -19 -7 ↓ - hot-/cold-smoked or gravlax Other preserved meat products 485 450 -7 -9 ↓ - other meat (excluding poultry) Ready meals consumed with further 251,193 234,288 -7 -3 ↓ heating, non-refrigerated Deli salads - with meat 52,485 51,035 -3 -5 ↓ Seite 12 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de 1.3 Quality of the underlying data set For the present evaluation of consumption trends for ready-to-eat foods, data on sales vol- umes and buyer households have been supplied by GfK SE. These data are based on a rep- resentative sample of private households in Germany (N = 30,000), stratified according to certain attributes such as German state (‘Land’), household size, age, number and age of children, social status, nationality, etc. The data can therefore be viewed as representative for the sum total of all German households. For each survey period, data for 12 months are consolidated. This also accounts for the seasonal variability in patterns of buyer behaviour. Since the panel data are retained for only five years, only this period of time can be ac- counted for by the evaluation. A 10-year retrospective analysis is therefore not possible. The survey method did not change between the two periods considered: as a result, the data rec- ords from both periods can be compared with one another. The survey method, using the scanning of the EAN barcode on the food product, enables an exact record of the purchase while minimising recall bias. Nonetheless, it is still possible that products without an EAN barcode (e.g. fresh fruit) in particular or products that are con- sumed immediately (e.g. ready meals that require no further heating) are underrepresented, as the data input is more complicated or the purchase was simply forgotten after consump- tion. Nor are any products reported on from out-of-home consumption (bakers, canteens, snack bars, etc.), since the ‘forget rate’ is too high, which makes the available data unrelia- ble. Accordingly, one may assume that the actual consumption in this convenience segment is underestimated. The underlying criterion for what is probably a more frequent rate of consumption consists of sales volume in tonnes and the number of buyer households in percent, which are compared in two periods spaced five years apart. A rise in sales volume with a simultaneously constant number of buyer households suggests that the purchasing households are either buying the products more frequently or that the quantity per purchase was increased. At the same time, however, this can also be influenced by retailers increasing the available pack contents. If figures for sales volume and buyer households rise simultaneously, one would expect an in- crease in acceptance by households, as can be observed with many products from the con- venience segment (ready meals, pre-cut fruit). This criterion permits the interpretation of a very high likelihood of more frequent consumption. However, this criterion does not exclude a scenario where products with increased sales volume and a constant or even lower number of buyer households are also being consumed in greater quantities. The sales volume for buffalo milk cheese has risen sharply, for example, although the number of buyer households has fallen. Since both sales volume and buyer households in the ‘Cheese and dairy products’ parent group have remained the same overall, one may suspect a substitution effect is at work here. Accordingly, households are actually buying the same quantity of cheese but some of this is now buffalo milk cheese. Once again, this would mean that buffalo milk cheese was being consumed more frequently. Overall, data considered at a higher level of aggregation show less movement than when data are considered for foods differentiated by type. In the ‘Fruit’ group, for example, virtually no change is seen, while the sub-group of ‘Mixed fruit salad/pre-cut fruit’ shows an increase in sales volume of 242%. As a result, movements within larger segments cannot be discovered on the basis of the available data. One group affected by this fact is ‘Other preserved meat products - beef’, which exhibits a high rate of growth, even though no statement can be made about the underlying cause. In- creases similar to those observed for fruit salad/pre-cut fruit can therefore be assumed even in more differentiated segments. Nor can ready meals, pastries or various items of dairy products be further differentiated, since neither the budget nor the level of detail provided for Seite 13 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de the product description permit a more differentiated analysis. The latter affects the microbio- logically relevant distinction applicable to raw milk products, cream-filled pastries and hot- or cold-smoked fish products. While the criterion of a parallel rise in sales volume and buyer households permits conclusions to be drawn about a strong possibility of more frequent con- sumption, it does not exclude scenarios where other products are being consumed more fre- quently with a rise only in sales volume or at a lower level of data aggregation. The present data are marketing data surveyed at the household level. As such, they do not permit any conclusions to be drawn about actual consumption quantities or frequencies, however, as no information is available about the further use of these products or food waste generated. Nor can any statements be made about subpopulations such as children, preg- nant women or older individuals, since the data were not surveyed at an individual level. Whether sensitive groups of individuals consume ready-to-eat foods with a greater or lesser frequency cannot be evaluated using the present data. 1.4 Conclusions To answer the question of whether ready-to-eat foods with the potential for contamination with Listeria monocytogenes were consumed with greater or lesser frequency, market data was consulted from the representative household panel from GfK SE. Since these data pro- vide no information about actual consumption quantities or frequencies, a criterion for an in- creased/reduced consumption was specified as the parallel decrease/increase in sales vol- ume and the number of buyer households as a comparison between two periods spaced five years apart. A total of 40 products were selected for the assessment. Of these, 14 show an increase in sales volume and buyer households. These include a variety of products from the food groups of fruit, vegetables, cheese, deli salads and ready meals. Products from the con- venience segment (‘Mixed fruit salads/pre-cut fruit’ and ‘Ready meals consumed without fur- ther heating’) in particular show the largest gains from a low starting-point. These gains are also in all likelihood underestimated, since no out-of-home consumption is accounted for and products from high street retail are often not reported (recall bias). Five of the 40 products show a decline in sales volume and buyer households, including various foods from the cate- gory ‘Other meat (excluding poultry)’ (e.g. game), hot- and cold-smoked or gravlax fish, ready meals (dried products) and deli salads containing meat. All remaining foods show only small changes or changes in opposite directions (e.g. growth in sales volume plus decline in buyer households). The products for which a parallel increase/decrease in sales volume and buyer households is recorded are in all probability also more frequently or less frequently consumed, respectively. This does not mean that products whose trends move in opposite directions are not in fact consumed more frequently or less frequently, merely that this state- ment cannot be made with an adequate level of confidence. Seite 14 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de 2. Response to the following question: Which types of ready-to-eat food (e.g. of ani- mal or plant origin) were most frequently found to be contaminated with Listeria monocytogenes during the last 10 years? 2.1 Underlying data In answering the question about the types of ready-to-eat food most frequently found to be contaminated with Listeria monocytogenes during the last 10 years, the annual trend reports from the BfR on pathogens for zoonoses in Germany were consulted as the most significant source of underlying data [5-12]. The basis for these reports is formed by the annual surveys of the results of investigations into zoonotic pathogens, which are carried out by state-level authorities in Germany as part of the monitoring of food, animals, feed and environmental samples. For the evaluation, the qualitative and quantitative results of investigations for Lis- teria monocytogenes were consulted from routine samples taken from 2008 to 2016. At the time when the answer to the official decree was being prepared, data for routine samples from 2017 had not yet been fully validated and were therefore not considered in the evalua- tion. As a supplement to the data from food monitoring, the results of zoonoses monitoring from the years 2012 to 2017 were also consulted [13-18]. In this context, qualitative and quantita- tive investigations relating to Listeria monocytogenes have been performed since 2012, pri- marily for product groups such as ready-to-eat foods, for which only sparse data were availa- ble. The investigation period for the individual food matrices covered a year in each case for zoonoses monitoring. Unlike the annual surveys performed by the food monitoring authori- ties, these data therefore do not reveal a trend extending over the past few years. Since rep- resentative sample sizes for the whole of Germany were typically investigated, however, these data can offer initial indications of ready-to-eat foods that are more frequently contami- nated with Listeria monocytogenes. In the context of an EU baseline study on the occurrence of Listeria monocytogenes in cer- tain ready-to-eat foods, prevalence data for smoked and gravlax fish, for heat-treated meat products, and for cheese made from cow’s milk were also investigated in 2010 and 2011. The results of this study for Germany were also included in the evaluation [19, 20]. To be able to compare the prevalence data on the occurrence of Listeria monocytogenes in ready-to-eat food with the consumption data evaluated for question 1, the surveyed product groups of ready-to-eat foods were examined, where possible, at the same and smallest-pos- sible level of aggregation for the first two questions of interest. The product groups consid- ered have been briefly outlined above as part of the answer to question 1 (see section 1.1, ‘Underlying data’). As a result of differences in the surveying and reporting of consumption data and prevalence data, however, the foods reported on are not identical across all product groups. Corresponding deviations are compared with one another in table 3. Seite 15 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de Table 3: Deviations in the grouping applied to obtain product groups and foods for the evalu- ation of prevalence data on Listeria monocytogenes in ready-to-eat foods and the consump- tion of ready-to-eat foods in Germany Product group/ Grouping used for the evalua- Grouping used for the evalua- food tion of prevalence data on Lis- tion of consumption data for teria monocytogenes in ready- ready-to-eat foods to-eat foods Heat-treated meat • Pork • Pork products • Beef • Beef • Poultry (overall) • Including poultry (chicken, tur- key) • Including poultry (other domes- tic poultry) • Other meat (excluding poultry) • Other meat (excluding poultry) Fish products • Fish, cold-smoked or gravlax • Fish, hot-/cold-smoked or grav- (marinated accordingly) lax • Fish, hot-smoked • Fish, other means of preserva- • Fish, marinated (not gravlax) tion (marinated accordingly, not gravlax) • Fried fish and boiled fish foods, • Fried fish and boiled fish foods, fish in aspic, long shelf-life fish fish in aspic, long shelf-life fish products products, pre-cooked prawns, shrimp, crayfish Cheese • Soft cheese • Soft cheese and semi-hard • Other cheese (incl. semi-hard cheese cheese) • Other cheese Minced meat prepa- • Minced meat preparations (any, • Steak tartare, pork mince with rations intended to incl. those for raw consumption seasoning be consumed raw such as steak tartare and pork mince with seasoning, as well as others not intended for raw consumption) Ready meals • Ready meals (any, with/without • Ready meals consumed with- further heating) out further heating • Ready meals consumed with further heating, non-refrigerated • Ready meals consumed with further heating, chilled • Ready meals consumed with further heating, frozen Seite 16 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de 2.2 Results With only a few exceptions, ready-to-eat foods can be found in almost all product groups of vegetable and animal origin. As a result of the widespread presence of Listeria monocyto- genes in the environment, and its high degree of tenacity in production environments in com- panies that manufacture and process food products, the pathogen can contaminate a wide variety of ready-to-eat foods prior to, during or after processing. In Germany, high rates of detection have been recorded over the last 10 years in routine samples taken from cold-smoked and gravlax fish (7-18%), hot-smoked fish (3-9%), fish with other means of preservation (4-10%) (table 10) and other preserved meat products (10-17%) (table 8). In all of these product groups, quantitative investigations have regularly determined values that exceed the safety criterion for Listeria monocytogenes as specified in EU Regula- tion (EC) No 2073/2005 for ready-to-eat foods (tables 9 and 11). In this analysis, concentra- tions of Listeria monocytogenes (a ‘viable colony count’ in subsequent sections) in cold- smoked and gravlax fish as well as hot-smoked fish were more likely to exceed 100 CFU/g (in 0.3-2.4% and 0.1-1.8% of quantitatively investigated samples) than in fish with other means of preservation and other preserved meat products (0.2-0.8% and 0.3-0.8% of quanti- tatively investigated samples). Viable colony counts exceeding even 1,000 CFU/g were de- tected in all product groups at isolated to regular frequencies. Differentiation of other pre- served meat products by animal species revealed slightly higher rates of detection for prod- ucts made from pork (8-13%) than for products made from the meat of other animals, exclud- ing poultry (5-10%). Listeria monocytogenes was also detected regularly in the category of other preserved meat products from beef. Since available data here are sparse, however, this prevents the derivation of valid statements about prevalence (table 8). High rates of prevalence for Listeria monocytogenes are also exhibited by minced meat (7- 22%) and minced meat preparations (15-24%) (table 4). In the minced meat preparations product group, prepared pork mince is offered by retailers to the consumer for raw consump- tion as a seasoned mince product, while raw beef mince is offered as a steak tartare or other minced steak product. These foods are therefore to be considered as ready-to-eat. However, the detection rates specified here relate to all minced meat preparations, including those that are intended to be heated before consumption. This therefore limits the derivation of conclu- sions from trend reporting data about certain minced meat preparations intended to be eaten raw - such as seasoned pork mince and steak tartare. During 2017 zoonoses monitoring, Lis- teria monocytogenes was detected in 11% of the samples of steak tartare/minced beef steak investigated (table 21). In routine samples of minced meat and minced meat preparations, viable colony counts ex- ceeding 100 CFU/g were detected regularly and even viable colony counts exceeding 1,000 CFU/g were detected frequently (Table 5). In contrast, the samples of steak tartare/minced beef steak investigated during 2017 zoonoses monitoring did not exceed the limit value of 100 CFU/g (Table 22). In recent years, lower detection rates of Listeria monocytogenes occurred in heat-treated meat products overall (1.0-3.3%) (Table 6), deli salads (1.0-4.1%) (Table 17), cheese from cow’s milk (0.0-2.2%) (Tables 12 and 13), pastries (0.4-1.4%) (Table 15) and ready meals (0.0-2.7%) (Table 18). In heat-treated meat products differentiated by animal species, some detection rates are higher than for heat-treated meat products overall (pork products 1.6-4.4%; products from other meat excluding poultry 1.1-6.5%; including poultry 1.5-5.6%) (Table 6). For heat- Seite 17 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de treated meat products made from beef, available data are sparse; for cheese, valid state- ments can be made only for cheese made from cow’s milk. Slightly higher rates of detection were recorded here for soft cheese made from raw milk (0.0-1.6%) and from pasteurised milk (0.0-2.2%) than for other cheese made from pasteurised milk (0.2-0.8%) (Tables 12 and 13). Values exceeding the safety criterion specified for ready-to-eat foods of 100 CFU/g were re- ported for isolated cases of deli salads, cheese made from cow’s milk, pastries and ready meals, as well as regularly for heat-treated meat products (0.1-0.2%) (Table 7). Viable colony counts even exceeding 1,000 CFU/g were reported in samples from all food groups men- tioned, excepting ready meals. Listeria monocytogenes has been detected sporadically in recent years in lettuce and vege- tables, fruit, and goat’s/sheep’s cheese and other raw milk cheese made from cow’s milk. However, the available data are inadequate for the purpose of deriving valid statements about prevalence for these product groups. 2.3 Individual results by product group The following section provides a detailed description of the results on the occurrence of Lis- teria monocytogenes in the individual product groups of ready-to-eat foods and summarises the relevant data from Tables 4 to 22 in the annex. Minced meat and minced meat preparations Over the last few years, qualitative investigations of routine samples of minced meat and minced meat preparations reveal constantly high detection rates for Listeria monocytogenes. For minced meat, 15-22% positive samples were reported in the years from 2008 to 2016 (median of annual sample size N = 1,070). One exception was the year 2013, with only 7% of samples being positive (Table 4). The results of quantitative investigations of minced meat reported in the same period revealed viable colony counts exceeding 100 CFU/g in 0.1-1.6% of samples. In seven out of nine years, viable colony counts exceeding 1,000 CFU/g were reported in 0.1-0.2% of samples (Table 5). For minced meat preparations, data are available for the years 2011 to 2016. For this period, a prevalence of 15-24% was determined (median of annual sample size N = 1,363). The via- ble colony counts reported here were over 100 CFU/g for 0.5-1.0% of the quantitatively in- vestigated samples and, in four out of six years, over 1,000 CFU/g in 0.1-0.3% of samples (Table 5). For routine samples reported for pork mince preparations, a significantly smaller data set is available for the years 2011 to 2016 (median of annual sample size N = 94). Lis- teria monocytogenes was detected here in 5-25% of the qualitatively investigated samples. In three out of six years, 0.5-1.5% of samples exhibited viable colony counts between 100 and 1,000 CFU/g. For steak tartare/minced beef steak, valid data are available only from zoonoses monitoring in 2017. Listeria monocytogenes was detected qualitatively in 11% of samples investigated (N = 278) (Table 21). Of the quantitatively investigated samples (N = 251), 2% exhibited Lis- teria monocytogenes at a level under 100 CFU/g. No sample exhibited a viable colony count over 100 CFU/g (Table 22). Heat-treated meat products For heat-treated meat products, qualitative investigations of routine samples taken from 2008 to 2016 reported 1.8-3.3% of samples positive for Listeria monocytogenes (median of annual Seite 18 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de sample size N = 2,367). One exception was the year 2016, with only 1.0% (Table 6). A com- parison differentiated by animal species revealed detection rates for heat-treated meat prod- ucts made from pork of 1.6-4.4% (median of annual sample size N = 731), for heat-treated meat products made from other meat excluding poultry of 1.1-6.5% (median of annual sam- ple size N = 614) and for heat-treated poultry meat products of 1.5-5.6% (median of annual sample size N = 368). A significantly smaller data set was available for heat-treated meat products made from beef (median of annual sample size N = 35). The detection rates for Lis- teria monocytogenes varied here between 0 and 8%. For quantitatively investigated heat-treated meat products, routine samples with viable col- ony counts exceeding 100 CFU/g (0.1-0.2%) were reported for all years with the exception of 2016. In four years, 0.1% of samples also exhibited viable colony counts exceeding 1,000 CFU/g (Table 7). For heat-treated poultry meat products, the viable colony counts reported in 0.3-0.9% of quantitatively investigated samples were more frequently above 100 CFU/g than in heat-treated meat products made with meat from other animals (0.2-0.5%). In heat-treated meat products made from beef, no viable colony counts were determined in excess of 100 CFU/g. As with the qualitative investigations, only a few samples were quantitatively investi- gated here (median of annual sample size N = 35). In the course of the 2010/2011 EU baseline study on Listeria monocytogenes in ready-to-eat foods, the pathogen was detected in 1.9% of the samples of heat-treated meat products in- vestigated in Germany (N = 915) (Table 21). Viable colony counts in excess of 100 CFU/g were exhibited by 0.1% of samples investigated (Table 22). Other preserved meat products During the years 2008 to 2016, qualitatively investigated routine samples of other preserved meat products (overall) in Germany exhibited high rates of contamination with Listeria mono- cytogenes amounting to 10-17% (median of annual sample size N = 2,392) (Table 8). During the same period, viable colony counts of over 100 CFU/g were determined in all years for 0.3-0.8% of quantitatively investigated routine samples. With the exception of 2010 and 2016, 0.1-0.3% of samples investigated also exhibited viable colony counts exceeding 1,000 CFU/g (Table 9). Other preserved meat products made from pork were more frequently contaminated, amounting to 8-13% of qualitatively positive samples (median of annual sample size N = 527) (Table 8). Viable colony counts exceeding 100 CFU/g were determined for 0.2-1.0% of quan- titatively investigated samples. In the period under investigation, this affected six out of nine years. In 2008 and 2014, viable colony counts exceeding 1,000 CFU/g were also determined in 0.3-0.5% of samples investigated (Table 9). For other preserved meat products made from beef, valid statements on the occurrence of Listeria monocytogenes are not possible, as a result of the limited availability of data from 2008 to 2016. In the few samples investigated annually, the pathogen was regularly de- tected, however. Viable colony counts exceeding 100 CFU/g were also determined. For other preserved meat products made from other types of meat excluding poultry (e.g. game), usable data are available only for the period 2010-2014. Listeria monocytogenes was detected here in 5-10% of qualitatively investigated routine samples (median of annual sam- ple size N = 488) (Table 8). During the same period, viable colony counts over 100 CFU/g were detected in three years in 0.2-0.3% of quantitatively investigated samples; in 2011, the viable colony count also exceeded 1,000 CFU/g in 0.2% of samples (Table 9). Seite 19 von 55
Bundesinstitut für Risikobewertung www.bfr.bund.de During the course of the 2017 zoonoses monitoring, spreadable raw sausage products from the ‘Other preserved meat products’ group were investigated for Listeria monocytogenes. The pathogen was detected in 12% of qualitatively investigated samples (N = 393) (Table 21). In the quantitative investigation (N = 378), two samples (0.5%) exhibited viable colony counts exceeding 100 CFU/g, amounting to 220 and 580 CFU/g (Table 22). Fish products In the product group of fish products, routine samples investigated for cold-smoked fish, gravlax fish and hot-smoked fish as well as samples for fish using other means of preserva- tion exhibited high levels of prevalence for Listeria monocytogenes on a continuous basis from 2008 to 2015. For cold-smoked and gravlax fish, 13-18% of qualitatively positive sam- ples were determined, with the exception of 2011 (8%) and 2015 (7%) (median of annual sample size N = 374) (Table 10). In addition, 0.3-2.4% of samples with viable colony counts over 100 CFU/g were reported consistently across all years. With the exception of 2010 and 2011, 0.3-0.8% of quantitatively investigated samples also exhibited higher viable colony counts exceeding 1,000 CFU/g (Table 11). For hot-smoked fish, a slightly lower prevalence of 3-9% was determined in qualitatively in- vestigated routine samples (median of annual sample size N = 873). With the exception of 2015, viable colony counts over 100 CFU/g were determined in 0.1-1.8% of quantitatively in- vestigated samples. In 2008 to 2012 and in 2016, viable colony counts exceeding 1,000 CFU/g were also detected in 0.2-1.0% of samples. Smoked and gravlax fish were investigated in the course of the EU baseline study in 2010/2011 on the occurrence of Listeria monocytogenes in certain ready-to-eat foods. No distinction was made in this study between hot- and cold-smoked fish. The pathogen was discovered in 12% of qualitatively investigated samples (N = 474) (Table 21). Of these, 1.5% of samples exhibited viable colony counts of over 100 CFU/g (Table 22). These results are consistent with the determined prevalence data as reported annually for routine samples. In the category of ‘Fish with other means of preservation’, lower prevalence was also de- tected here (4-10%) compared with cold-smoked and gravlax fish (Table 10). With the excep- tion of 2010, quantitative investigations revealed viable colony counts over 100 CFU/g in 0.2- 0.8% of routine samples. In isolated years, 0.1-0.4% of samples exhibited viable colony counts exceeding 1,000 CFU/g (Table 11). For routine samples of heat-treated fish, only a small data set was reported that does not permit valid statements to be made about the occurrence of Listeria monocytogenes. Iso- lated positive samples were nonetheless discovered. Quantitatively investigated samples ex- hibited viable colony counts under 100 CFU/g. Cheese Prevalence data for Listeria monocytogenes in routine samples of cheese from 2008 to 2016 were grouped into categories according to cheese type (soft cheese, other cheese excluding soft cheese), the milk processing condition (raw milk, pasteurised milk) and milk origin (ani- mal species). Samples of cheese made from cow’s milk made up the majority of the samples investigated here. With the exception of the year 2012 (0% with N = 91), a prevalence of 0.4- 1.6% was determined for raw milk soft cheese made from cow’s milk (median of annual sam- ple size N = 192) (Table 12). In soft cheese made from pasteurised cow’s milk, 0.3-1.0% positive samples were detected with only a few exceptions (median of annual sample size N = 601). In 2009 and 2014, a slightly higher level of prevalence was determined, at 2.2% (Ta- ble 13). Other cheese from pasteurised cow’s milk exhibited levels of prevalence from 0.2- Seite 20 von 55
You can also read